BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Asia/Seoul
X-LIC-LOCATION:Asia/Seoul
BEGIN:STANDARD
TZOFFSETFROM:+0900
TZOFFSETTO:+0900
TZNAME:KST
DTSTART:18871231T000000
DTSTART:19881009T020000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20230103T035309Z
LOCATION:Room 325-AB\, Level 3\, West Wing
DTSTART;TZID=Asia/Seoul:20221207T140000
DTEND;TZID=Asia/Seoul:20221207T153000
UID:siggraphasia_SIGGRAPH Asia 2022_sess162_papers_262@linklings.com
SUMMARY:PADL: Language-Directed Physics-Based Character Control
DESCRIPTION:Technical Communications, Technical Papers\n\nPADL: Language-D
 irected Physics-Based Character Control\n\nJuravsky, Guo, Fidler, Peng\n\n
 Developing systems that can synthesize natural and life-like motions for s
 imulated characters has long been a focus for computer animation. But in o
 rder for these systems to be useful for downstream applications, they need
  not only produce high-quality motions, but must also provide an accessibl
 e and versatile interface through which users can direct a character's beh
 aviors. Natural language provides a simple-to-use and expressive medium fo
 r specifying a user's intent. Recent breakthroughs in natural language pro
 cessing (NLP) have demonstrated effective use of language-based interfaces
  for applications such as image generation and program synthesis. In this 
 work, we present PADL, which leverages recent innovations in NLP in order 
 to take steps towards developing language-directed controllers for physics
 -based character animation. PADL allows users to issue natural language co
 mmands for specifying both high-level tasks and low-level skills that a ch
 aracter should perform. We present an adversarial imitation learning appro
 ach for training policies to map high-level language commands to low-level
  controls that enable a character to perform the desired task and skill sp
 ecified by a user's commands. Furthermore, we propose a multi-task aggrega
 tion method that leverages a language-based multiple-choice question-answe
 ring approach to determine high-level task objectives from a language comm
 and. We show that our framework can be applied to effectively direct a sim
 ulated humanoid character to perform a diverse array of complex motor skil
 ls.\n\nRegistration Category: FULL ACCESS, ON-DEMAND ACCESS\n\nLanguage: E
 NGLISH\n\nFormat: IN-PERSON, ON-DEMAND
URL:https://sa2022.siggraph.org/en/full-program/?id=papers_262&sess=sess16
 2
END:VEVENT
END:VCALENDAR